ROS2 Introspection

Execute core ROS 2 introspection commands to query the ROS graph (topics, nodes, services, actions, parameters). STRICTLY read-only.

This skill appears to do exactly what it claims: run read-only ROS 2 introspection commands. Before installing or invoking it, 1) run the provided scripts locally and inspect them (they are short and readable); 2) run setup.sh yourself to confirm the ros_setup_path it writes; 3) be cautious when running commands like 'topic echo' or 'param dump' since they can display or stream topic/parameter values that may contain sensitive robot data; 4) note that 'rqt_graph' is executed as a special case (it will launch a GUI) and is not passed through the same validation flow — if you don't want GUIs launched, avoid that subcommand; and 5) run the skill in an environment you control (or a container) if you have sensitive topics or networks. Overall the package is internally consistent and low-risk provided you follow these precautions.

Type: OpenClaw Skill Name: ros2-introspection Version: 1.0.1 The skill provides a well-structured and security-conscious interface for ROS 2 introspection. It implements a Python wrapper (safe_ros2_introspection.py) that enforces a strict allowlist of read-only commands and utilizes safe subprocess execution (shell=False) to mitigate command injection risks. The setup script (setup.sh) and configuration logic are transparent and align with standard ROS 2 environment requirements without any indicators of malicious intent or data exfiltration.